World Bank Pitches 'Small AI' as a Tool to Narrow Global Inequality

As governments and regulators around the world grapple with how to oversee artificial intelligence, a parallel debate is taking shape over whether the technology can serve as an engine for development and humanitarian aid rather than a source of deeper division. Warnings that AI could widen the gap between rich and poor nations have become commonplace, but the World Bank is making the case that the technology, if adapted properly, could also be harnessed for social good.

The discussion formed the focus of an episode of France 24's business programme "People & Profit," which examined AI's potential role in reducing inequality. The segment weighed the risks of the technology exacerbating existing disparities against the argument that it can be deliberately shaped to benefit developing economies.

At the centre of that argument is World Bank President David Malpass's successor in the institution's leadership, David Malpass having been replaced by Ajay Banga, though the source material focused on the perspective of a senior World Bank figure advocating for a tailored approach. According to the programme, the institution believes developing economies stand to gain from AI provided it is designed around their specific conditions rather than imported wholesale from advanced markets.

That approach is described as "small AI" — a term used to capture technology that is more closely adjusted to local circumstances, can run offline, and is capable of operating in environments where internet connectivity or electricity supply may be unreliable or limited. Rather than relying on massive, cloud-dependent systems that demand constant power and high-bandwidth connections, small AI is envisioned as lighter, more resilient, and more relevant to the daily realities of poorer regions.

The reasoning is straightforward: many of the communities that could benefit most from AI-driven tools for health, agriculture, or education are precisely those least able to support the infrastructure that mainstream AI systems require. A model that functions with intermittent power and poor connectivity would therefore remove one of the most significant barriers to adoption.

The World Bank's position also stresses that the rapid transformation of daily life through technology should not distract development agencies from their core mission. According to the argument laid out in the programme, institutions must remain focused on tackling inequality and meeting basic human needs, including access to clean water, healthcare, and education.

The point is not to treat AI as a substitute for those fundamentals, but to ensure that the excitement surrounding the technology does not crowd out the long-standing priorities of development work. In that view, AI should be understood as a tool that sits alongside traditional investment in infrastructure and public services, not as a shortcut that replaces them.

Crucially, the World Bank's message is not one of unqualified optimism. The institution acknowledges that AI brings challenges as well as opportunities, and that its benefits are not automatic. The difference between a technology that narrows inequality and one that deepens it, according to this framing, depends heavily on the policies that surround it.

That emphasis on policy is central to the "small AI" concept. The argument holds that tailored AI, combined with the right regulatory and governance frameworks, can play a positive role in development. Without such guardrails, the same tools risk reinforcing the advantages of wealthier countries and better-connected communities.

The debate over AI and inequality has intensified as the technology spreads across sectors, from finance and manufacturing to public administration and humanitarian response. Critics have warned that the concentration of AI research and computing power in a handful of wealthy nations could leave developing countries as consumers of foreign technology rather than participants in its creation.

The World Bank's intervention suggests a different path, one in which AI is localised and adapted rather than simply exported. By emphasising offline capability and lower infrastructure demands, the "small AI" approach implicitly challenges the assumption that cutting-edge technology must always be resource-intensive.

The discussion also reflects a broader shift in how development institutions talk about innovation. Rather than presenting new technology as a cure-all, the emphasis is on integration — ensuring that AI complements efforts to expand access to water, health services, and schooling, which remain the foundation of any long-term reduction in poverty.

For developing economies, the stakes are considerable. If AI can be tailored to local needs, it could improve agricultural yields, extend the reach of medical diagnostics, and make educational resources available in remote areas. If it cannot, the risk is that the digital divide becomes an intelligence divide, with poorer nations falling further behind.

The programme's framing suggests that the outcome is not predetermined. The choice between AI as a driver of inequality and AI as a tool for inclusion rests on decisions made by governments, development agencies, and technology companies in the coming years.

The World Bank's argument ultimately rests on a balance: recognising the genuine transformation underway while refusing to lose sight of the basics. As the institution sees it, small AI, harnessed correctly and supported by sound policy, can be part of the solution rather than part of the problem — provided the world does not allow the promise of new technology to obscure the enduring needs it is meant to serve.